Coevolution Paper figs

This Study

To better understand the dynamic interaction between spatial structure, genetic architecture, and coevolution, we conducted a simulation study, exploring a range of situations plausible for the Taricha newt Thamnophis garter snake system.

To better understand the process of adaptive evolution across geographical space, this paper asks three main questions:

How does the genetic architecture of the traits in newts and snakes affect how they coevolve?
Under what situations do we get spatial patterns of correlated traits as we see in the real world?
How fast does coevolution increase resistance and toxicity in these organisms with different combinations of genetic architectures?

In particular, we compare different levels of mutational variance and polygenicity using individual-based simulations of continuous geographic space. The results complement field observations by describing situations that are consistent with empirical observations, and exploring other possible outcomes.

The experiments

Experiment 1: varies both mutation rates and mutation effect sizes. Experiment 2: mutation rate was fixed for both species in each simulation, but allows the species to have different mutation effect sizes (and hence mutational variance). Experiment 3: mutational variance was the same for newts and snakes in each simulation, although polygenicity could be different (by varying mutation rate and mutation effect size).

How to use this file?

run all of the simulations
run gather_data.Rmd to make lit (phenotypes, population size, …), cor (correlations), and grid (values calculated from grids) file
all files will be read in through the Coevolution folder
All text files used to generate this markdown are included in the Coeovlution folder.

Functions

find_coevo_slope = calculating the slope of coevolution
coevo_slope_flat_map_snake_change = making the figure for snake phenotype change
coevo_slope_flat_map_newt_change = making the figure for newt phenotype change
boxplot_coevo_slope_small_exp_color = showing how phenotypes change for normal, no heritability, and random interaction outcome
fast_slow_no_coevolution = showing an example of co-evolution of phenotypes throughout time

Experiment 1 Flatmap

Varies both mutation rates and mutation effect sizes. We ran the simulation normally, collecting data on how newt toxicity and snake resistance changed throughout the course of the simulation. We also ran the simulation without heritability and without the newt-snake interaction. All the following results are from a flat map.

Speeds of Coevolution

Change in Newt Phenotype by Different Genetic Architectures

Change in Snake Phenotype by Different Genetic Architectures

Change in Newt and Snake Phenotype

Spatial Phenotype Correlation for Experiment 1

What is the correlation between newt toxicity and snake resistance across geographical space within the simulation? Simulation varies both mutation rates and mutation effect sizes (Experiment 1). In these simulations there were different maps that created heterogeneous landscapes (cost gradient, interaction rate gradient).

Spatial phenotype correlation FlatMap/CostMap/InteractionMap

Spatial phenotype correlation NoInteractionCostMap/SnakeCostMap/NewtCostMap

Spatial phenotype correlation overtime FlatMap/CostMap/InteractionMap

Reading in the grid files for the cost gradient

What do local newt and snake phenotypes look like within the simulation? A few simulations from Experiment 1 (varies both mutation rates and mutation effect sizes) with a cost heterogeneous landscapes.

Mean Phenotypes for the Cost Gradent: Snake 4a, Newt 3a

Min Phenotypes for the Cost Gradent: Snake 4a, Newt 3a

Min Phenotypes for the Cost Gradent: Snake 2a, Newt 4a

Min Phenotypes for the Cost Gradent: Snake 4a, Newt 2a

Mean Phenotypes for the Cost Gradent: Snake 2a, Newt 4a

Mean Phenotypes for the Cost Gradent: Snake 4a, Newt 2a

Speed of Coevolution (experiment 2 and experiment 3)

Does the speed of coevolution depends more on mutational variance than it does on polygenicity? Here we plot the Coevolution Speed (Phenotype Change/ Time), with results from experiment 2 and 3.

Experiment 2: mutation rate was fixed for both species in each simulation, but allows the species to have different mutation effect sizes (and hence mutational variance). Experiment 3: mutational variance was the same for newts and snakes in each simulation, although polygenicity could be different (by varying mutation rate and mutation effect size).

Relationship between population size and phenotype

What is the relationship between population size and phenotype? This section examines results from experiment 1, 2, and 3 (from the early and late part of the simulation).

PopSize and Phenotype Experiment 1 Simulation

PopSize and Phenotype Experiment 2b Simulation

PopSize and Phenotype Experiment 2c Simulation

PopSize and Phenotype Experiment 2d Simulation

PopSize and Phenotype Experiment 2e Simulation

PopSize and Phenotype Experiment 3f Simulation

PopSize and Phenotype Experiment 3g Simulation

PopSize and Phenotype Experiment 3h Simulation

PopSize and Phenotype Experiment 3i Simulation

Plotting the relationship between phenotype/population size and GA

What is the relationship between phenotype/population size and GA? This section examines results from experiment 1, 2, and 3.